Systematic physicochemical characterization, carbon balance and cost of production analyses of activated carbons derived from (Co)-HTC of coal discards and sewage sludge for hydrogen storage applications

Hydrothermal carbonization (HTC) technologies for producing value-added carbonaceous material (hydrochar) from coal waste and sewage sludge (SS) waste might be a long-term recycling strategy for hydrogen storage applications, cutting disposal costs and solving waste disposal difficulties. In this study, hydrochars (HC) with high carbon content were produced using a combination of optimal HTC (HTC and Co-HTC) and chemical activation of coal tailings (CT), coal slurry (CS), and a mixture of coal discard and sewage sludge (CB). At 850 °C and 800 °C, respectively, with a KOH/HC ratio of 4:1 and a residence time of 135 min, activated carbons (ACs) with the highest Brunauer–Emmett–Teller specific surface (SBET) of 2299.25 m2g− 1 and 2243.57 m2g− 1 were obtained. The hydrogen adsorption capability of the produced ACs was further studied using gas adsorption isotherms at 77 K. At 35 bars, the values of hydrogen adsorbed onto AC-HCT (AC obtained from HTC of CT), AC-HCS (AC obtained from HTC of CS), and AC-HCB (AC obtained from HTC of the blending of coal discard (CD) and SS) were approximately 6.12%, 6.8%, and 6.57% in weight, respectively. Furthermore, the cost of producing synthetic ACs for hydrogen storage is equivalent to the cost of commercial carbons. Furthermore, the high proportion of carbon retained (>70%) in ACs synthesized by HTC from CD and SS precursors should restrict their potential carbon emissions.


Introduction
Carbon technologies have grown rapidly in the last two decades, owing to the introduction of new synthetic techniques such as pyrolysis, high-voltage-arc discharge, chemical vapour deposition, HTC, and laser ablation, as well as the growing demand for carbon-neutral alternatives due to environmental concerns [1]. South African coal beneficiation process plants generate around 65 Mt of coal waste per year, most of which is disposed of in tailing stockpiles and slurry dams [2]. South Africa, on the other hand, has roughly 824 wastewater treatment plants spread across 152 municipalities, with a total design capacity of 6.5 billion litres of wastewater/day, resulting in vast amounts of sewage sludge (SS) [3]. Numerous recent studies have revealed that due to a lack of feasible practical alternatives, SS disposal and piling have become the usual management choice for many wastewater treatment plants across the country [4][5][6].
On the other hand, the environmental and health concerns inherent to the establishment of coal waste and SS stockpiles are significant [7]. Reusing these wastes to create advanced, high-quality materials will reduce waste disposal costs. Furthermore, this will benefit society and mining companies by expanding new markets, creating green jobs, and lowering the carbon footprint. To achieve the above, novel and cost-effective waste reuse methods are required, such as the production of activated carbons (AC). Although AC has been widely synthesized from a variety of carbon precursors (petroleum coke, empty fruit bunch, bean husk, wood, palm solid waste, coconut shell, bituminous coals, and coal char), the porous properties of the resulting ACs are dependent on the precursors and the activation techniques used [8]. Furthermore, AC's physicochemical qualities influence its end application [9]. Several studies have also shown that the HTC process has the potential to be environmentally benign and cost-effective in the synergetic treatment of wet feedstock such as coal ash (CD) and SS to produce porous materials with high carbon content [10,11]. HTC combined with chemical activation, in particular, can improve the texture and porosity of the resulting AC. HTC combined with chemical activation, in particular, can improve the texture and porosity of the resulting AC.
Despite widespread recognition of hydrogen and fuel cell technology as potential future energy solutions, demonstrated pathways to the so-called "hydrogen economy", in which energy is stored and distributed as hydrogen, remain limited [12]. Among the technologies being investigated for hydrogen storage development are high-pressure gas, liquid hydrogen, complex hybrids, hydrogen intercalation in metals, and hydrogen adsorption on porous materials [13]. Numerous studies of the prospective transition from conventional to renewable fuels have shown that hydrogen has great economic potential as a future energy carrier. However, there are still barriers to widespread hydrogen use, the most significant of which is a lack of suitable storage solutions. Several storage strategies have been examined, with porous carbon materials, such as ACs, having a higher hydrogen absorption capacity than metal hybrids and complex hydrides [14,15]. Furthermore, the HTC process has been demonstrated to be economically viable, with a biochar (BC) production cost of 0.5 USD/kg, which is slightly less than that of the conventional pyrolysis technique (0.2 USD/kg) [16]. However, more research on the economics of producing ACs from waste for adsorptive hydrogen storage and the carbon balance of such systems is needed before such systems can be scaled up to become a more sustainable alternative to existing commercial options.
Therefore, the purpose of this study was to evaluate the use of AC obtained from HTC and Co-HTC treatment of CD and SS wastes for adsorptive hydrogen storage, as well as the associated physicochemical characteristics and production cost. To establish the financial viability and carbon footprint of coupling HTC and chemical activation processes to produce finely tuned ACs for hydrogen storage, the link between the physicochemical properties of the synthesized AC and the H 2 storage efficiency, as well as the economics and carbon emissions of producing ACs from CDs and SS, must be thoroughly investigated, and this paper provides a stepping stone in that direction.

Materials
For this study, coal tailing (CT) and coal slurry (CS) were collected from a coal beneficiation plant in Mpumalanga, South Africa. The obtained samples were dried in ambient air before being stored and transported in airtight bags to the laboratory, where they were dried once more at room temperature. The dried individual as-received samples were pulverized to −212 µm and -1 mm fractions in accordance with the ASTM D-2013 [17], with the physicochemical analyses (proximate, ultimate, total sulphur, scanning electron microscopy (SEM) and Fourier transform infrared (FTIR) spectroscopy) conducted on the −212 µm fraction. The SS used in this study was collected from a wastewater treatment plant (Ekurhuleni Water Care Company-ERWAT) and split into two portions. The first portion was dried for 24 h in a laboratory drier at 105 °C and crushed for physiochemical characterization, whereas the second portion was used for HTC experiments as received. The hydrothermal experiment was carried out in a high-pressure (110 mL) tubular reactor with a −1 mm coal fraction. The hydrothermally treated carbon material was chemically activated using various ratios of analytic-grade potassium hydroxide (KOH). The porous material generated from the activation process was neutralized using hydrochloric acid (HCl, 0.5N). Throughout the washing procedure of the neutralized products, distilled water was used. The chemical activation experiments were carried out in a tube furnace at various temperatures with a flow rate of 150 mL/min of nitrogen gas. Hydrogen gas baseline N5.0 from AFROX (African Oxygen Limited Company, South Africa) was used to perform the hydrogen adsorption tests on the synthesized ACs. 25 g of raw material (CT, CS, mixture of CT, CS and SS) was mixed with water as the input solvent in a solid liquid ratio of 1:4 in the tubular reactor. The HTC and Co-HTC process parameters (temperature and pressure) were varied to obtain the highest FC content carbonaceous material. The Co-HTC experiments used a feedstock mixture coal discard and sewage sludge (CB) consisting of 80% (CT+CS) and 20% SS blend. The CDs were mixed at a 1:1 ratio for experimental simplicity. In the HTC experiment, the reactor was heated to the desired reaction temperature and held for 95 min residence time. However, all Co-HTC experiments were performed with a 360 min residence time. The solid products obtained from the mix were identified as HC CT, HC coal slurry (HCS) and HC blend CD and SS (HCB). The obtained HCs were dried, characterized and prepared for chemical activation. AC synthesis was carried out from the HCs via a chemical activation process using potassium hydroxide. The chemical activation tests were performed at temperatures between 800 °C and 850 °C, with a heating rate of 15 °C/min. The holding time was varied from 90 min to 180 min before cooling using nitrogen. KOH/HC ratios of 1:1, 5:2 and 4:1 were used to optimize the chemical activation process in accordance with previous studies [17,18]. The produced ACs were washed with 0.5N hydrochloric acid (HCl) for residual alkali removal, and then with distilled water using a centrifuge until the mixture (product/distilled water) reached a pH around 6. The washed solid product from the mixture was dried at 60 °C in an oven for 24 h, prepared according to the ASTMD5142 standards for characterization and kept in an airtight container for further analysis. The optimized parameters obtained were used for direct chemical activation of CT and CS for the evaluation of the Brunauer-Emmett-Teller specific surface (S BET ). The specific surface area obtained was then compared with the S BET of AC produced from the combination of HTC and chemical activation. Afterwards, the optimized ACs produced were characterized and prepared for the hydrogen storage test. The chemical activation process methodology and application of the synthesized products are presented in Fig. 1.

Chemical activation
The activation experimental test runs were designed using Design-Expert software (DoE) (Version 13, State-Ease, Inc., Minneapolis, MN, USA). In the design, two parameters of the chemical activation process, namely temperature and residence time, were considered. For the experimental design, the central composite design (CCD) was utilized to determine the number of runs that will be required to optimize the process. Chemical activation experiments were conducted using temperatures between 800 °C and 850 °C and retention times between 90 min and 180 min. Based on the CCD, 13 experimental runs were performed for the optimization of the chemical activation of HCT, HCS and HCB. Five center points, four factorial points and four axial points were repeated five times to ensure reproducibility of the responses and to determine accurate experimental error. The response surface methodology (RSM) was adopted to understand the interactions of chemical activation factors on the S BET of AC produced from both HC samples. The AC with the highest S BET was selected as the optimized product. In addition, RSM was used to develop appropriate models to enable the prediction of the optimum operating conditions required to produce high S BET . To assess the significance of the designed models, analysis of variance (ANOVA) was performed. The numerical optimum response methodology option was then used to select the best chemical activation process conditions after using the maximum S BET as the optimization criteria for the produced ACs.

Characterization of the produced HCs and ACs
Proximate analysis of 1 g of −212 µm of raw materials (CT, CS, SS, CB), the produced HCs and ACs were carried out using thermogravimetric analyser equipment (Leco TGA 701 from LECO Corporation, St Joseph, MI, USA) in accordance with ASTM D5142 [19]. The determination of the elemental composition CHN (carbon, hydrogen, nitrogen) was done in accordance with ISO 12902 [17] standard method using a Flash 2000 Organic Elemental Analyzer (Thermo Scientific, Waltham, MA, USA), whereas oxygen was calculated separately as the difference from 100% using Eq. (1).
Equation (2) was used to determine the HTC and Co-HTC carbonization yields (Cy) to assess the degree of carbonization.
The produced ACs mass yield was calculated using Eq. (2).
Qualitative determination of the mineral phases in the produced ACs was conducted using D2 PHASER Bruker. The XRD instrument employed Cu-Kα radiation as the excitation source over a 2θ range, and a generator setting of 30 kV and 20 mA. The resulting diffractograms from the XRD analysis were matched on the Bruker D2 mineral phase database to identify the major mineral phases in the samples. The mineral phases were identified using X'PERT High-Score Plus analysis software. The XDR spectra obtained were compared with the spectra of the CD, SS and HCs to evaluate the transformation of the solid structure caused by the chemical activation process.
The FTIR analysis of the produced ACs was conducted on a Perkin Elmer spectrometer coupled with a diamond attenuated total reflectance (ATR) accessory in the wavenumber range between 4000 cm −1 and 450 cm −1 . The interpretation of the spectra was performed based on reference tables provided by Smith [20] to characterize various surface functionalities present in AC samples.
The SEM was used to observe the surface morphology of the produced ACs. The SEM was done using Carl Zeiss Sigma Field Scanning Electron Microscope connected to the Oxford X-act Energy Dispersive X-ray Spectroscopy (EDS) detector to obtain an elemental analysis. The SEM/EDS analysis settings were 10 kV and a working distance (WD) of 7.2-8.2 mm using a backscattered electron (BSE) signal.
The specific surface area and average pore diameter of the produced ACs were calculated using nitrogen (N 2 ) adsorption test at 77 K. The N 2 adsorption data were used to determine the S BET , micropore external surface area and pore size distribution (PSD) plot. For this analysis, the autosorb iQ gas sorption instrument (Quantachrome Instruments, Anton Paar, Boynton Beach, FL, USA) was used, while the Quantachrome ® ASiQwinTM software, which interfaces the autosorb iQ to a computer, was used for data acquisition and data reduction. Information on the distribution of micropores, mesopores and macropores was obtained from adsorption data of the N 2 isotherm using the nonlocal density functional theory (NLDFT) method.

Hydrogen adsorption
The hydrogen adsorption measurements were conducted using pure hydrogen in an ultra-clean vacuum system with all metal seals with a diaphragm and high-rate vacuum pumps. The density of hydrogen adsorbed was measured by the maximum amount adsorbed obtained from the isotherms compared to the total pore volume of the AC. The hydrogen adsorption capacity was measured according to the ASTM D 4607-94 [21] specifications. Isotherms of hydrogen adsorption on the ACs were measured using the high-pressure volumetric method at 77 K. The automatic high-pressure volumetric analyser (HPVA) from Micromeritics Particulate Systems coupled with a single-stage closed-cycle cryogenic refrigerator was used to perform hydrogen adsorption tests. Before hydrogen adsorption tests, the ACs were degassed in the vacuum system at 373 K overnight. Thereafter, the ACs were moved to a second degassing system into the HPVA in a 10 cm 3 cell filled with approximately 1.2 g of ACs at 373 K under vacuum for another 24 h. The pressure range of adsorption was fixed from 1 bar to 35 bar, and the obtained pressures of desorption were from 28 bar to 1.6 bar. The determination of hydrogen adsorption isotherms was carried out from cold and warm volume measurement after adsorption-desorption tests to avoid helium capture in narrow pores [22]. Tables 1 and 2 show the proximate and ultimate results of the raw materials and produced HCs. These findings indicate that the reaction temperature and pressure had a substantial influence on the HTC and Co-HTC processes. The maximum fixed carbon (FC) and total carbon (C) values were obtained for the HTC and Co-HTC processes at 150 °C and 210 °C, respectively. However, raising the temperature above 150 °C and 210 °C for HTC and Co-HTC, respectively, increases the ash content of the produced HC, probably due to the condensation of the material dissolved in the solid HCs and possible liquefaction of carbon from the feedstock. The mass yields of Co-HTC decreased more than those of HTC probably as a result of the greater degree of thermolytic degradation of cellulose and lignin in the SS under Co-HTC reaction conditions [22].

Characterization of the produced HCs
The results of proximate analysis showed that CS has a higher total carbon content than CT and CB. Using the ultimate analysis results and Eq. 2, the HTC process yield (Cy) was calculated and found to be 113.58%, 102.47% and 129.61% for HCT, HCS and HCB, respectively. This is consistent with the findings of Saba et al. [23], who studied the effect of biomass blends with coal on the increase in carbonization yield during the Co-HTC process. The investigation demonstrated that the increase in acidity caused by the degradation of biomass (SS) during the Co-HTC process enhanced the removal of inorganic elements from the feedstock. For both coal and SS blend, the oxygen level decreased, while the hydrogen amount varied slightly. This can be attributed to the hydrolysis and decarboxylation reaction occurring during HTC and Co-HTC [24]. As established by earlier investigations [24,25], the decrease in sulphur concentration in the produced HCs likely contributes to the formation of pores and improvement of the textural structure. Nitrogen gas used to keep an inert environment in the HTC reactor may have contributed to the modest increase in nitrogen in the generated HCs.
The results presented in Table 2 show that the increase in pressure to 27 bar and 22.5 bar led to an increase in the FC content of the HTC, HCS and HCB produced. The HCS with the highest FC content was selected for further analysis and chemical activation experiments. The diffractogram presented in Fig. 2a illustrates the findings of XRD investigation of coal samples, SS and produced HCs. The mineral phases identified and labelled accordingly on the diffraction patterns presented in Fig. 3 are  showing the trigonal, hexagonal, triclinic and monoclinic crystal systems of the three materials, respectively [26]. The correspondence of the XRD profiles of raw materials and produced HCs observed in Fig. 2a agrees with the reported results in the previous study conducted by Jiang et al. [27] on the characterization of coal/biomass blend. The decrease in peak intensities of mineral phases observed in the diffraction patterns of the produced HCs was proportional to the impact of HTC and Co-HTC operating conditions on the dissolution of mineral content of the feedstock. Figure 2a shows a significant decrease of the mineral phase peak's intensities for the Co-HTC process, probably favoured by the strengthening of the acidic medium produced by the decomposition of SS into organic acid monomers during the decarboxylation and dehydration reactions occurring in the HTC and Co-HTC processes. In addition, the XRD diffractogram indicates that the condensation and polymerization of carboxyl and carbonyl groups produced from decarboxylation during HTC and Co-HTC forms a porous hydrophobic aromatic polymer [28]. The FTIR spectra in Fig. 2b were used to describe the chemical structural properties of the raw materials and produced HCs. The interpretation of the spectra was performed based on reference tables provided by Smith [20]. Therefore, the peaks at the 3620 cm −1 band were attributed to the O-H stretching vibration linked with the kaolinite mineral phase presented on the XRD patterns in the form of silanol (Si-OH). The lower intensity of the O-H observed on the spectrum band of the produced HCs compared to untreated coals demonstrated evidence of dehydration reactions during the HTC and Co-HTC process. The intense peak at 2922.63 cm −1 was due to the higher presence of the aliphatic group in the SS characterized by long linear chains -CH 2 and -CH 3 asymmetric and symmetric vibrations (alkanes), respectively [29]. The shoulder peaks at the 2960 cm −1 band observed on the HCT and HCB spectra were assigned to asymmetric aliphatic -CH 3 stretching vibration produced from the interaction of the C-H aliphatic stretch observed at 2850 cm −1 and 2920 cm −1 and C-H alkyl groups from the CT and CS spectra during the process [30]. The peak observed at 1700 cm −1 can be assigned to the aliphatic C=O and -COOH stretching vibrations of carboxyl and carbonyls, mainly ketones, aldehydes, and esters in the  [29]. The low intensity of the C=O peaks in the produced HCs confirmed the occurrence of decarboxylation reactions during HTC and Co-HTC. The transmittance peak within the fingerprint region of the spectrum at 1500 cm −1 and 570 cm −1 band in all spectra with lowering intensity order from CT, SS, CS, CB, HCS, HCT to HCB were assigned to the vibrations associated with -CH 3 , C-C and Si-O-Si in quartz and kaolinite; aromatic species in aromatic rings; trans-and cis-CH 2 in long saturated aromatic -CH-CH chains and C-O stretching vibration of ether groups; and O-H bending vibrations in the phenolic, phenoxy and hydroxybenzene structures. The condensation and polymerization steps were confirmed by the decrease in peak intensity of the produced HCs compared to the raw materials, resulting in the elimination of oxygen and hydrogen into water in liquid phases and restructuration of the carbon skeleton [31]. The FTIR spectra revealed the complexity of the thermal decomposition and restructuration of the HTC and Co-HTC products. Figure 2 depicts the transition of raw material mineral phases and functional groups through the HTC and Co-HTC processes. The intensities of the detected peaks in the XRD patterns and FTIR spectra demonstrated the thermal process's influence on the feedstocks. The greater the drop in mineral phase intensities is, the greater the transmittance of the mineral matter functional group, corresponding to low adsorption intensity. This demonstrated that the HTC and Co-HTC processes reduced the mineral matter concentration of the raw material while increasing the amorphous shape of the generated HCs with a carbonaceous structural preponderance (lower transmittance percentage of carbonyl, aliphatic C-C and aromatic stretch). Consequently, the activation agent's interaction with the carbon of the generated HCs followed the carbon distribution from HCT, HCS and HCB. As a result, the greatest carbon content HC had the largest specific area and pore volume, which are related to the hydrogen adsorption capacity of the synthesized ACs.
The analysis of the obtained HCs indicates that the HTC and Co-HTC processes increased the carbon content of the feedstocks. The physicochemical parameters of the HCs synthesized in this investigation were comparable to those of prior studies [23,32,33]. Furthermore, the findings of this investigation show that the Co-HTC process has a considerable carbonization capability for producing HC from the combination of coal waste and SS (HCB), with comparable properties to HC generated from other feedstock, as described in earlier studies [23,34]. Furthermore, when compared to the reported operating conditions for Co-HTC and other thermal processes, such as pyrolysis and gasification [24], the optimal HTC and Co-HTC conditions derived in this study have the potential to reduce the energy requirement and cost of the process from a development and economic feasibility standpoint.

Chemical activation results
The statistical parameters were examined using ANOVA on the BET (Brunauer-Emmett-Teller) analysis results. The ANOVA results indicate that an interaction effect of process temperature and holding time had a significant influence on the S BET of synthesized ACs. For the obtained responses, the regression models (Eqs. 4-6) were evaluated using the probability (p value) and Fisher test values (F value). Furthermore, the developed regression models (Eqs. 4-6) confirmed the influence of the two process variables, namely temperature and holding time, on the S BET of the produced ACs. Because of their high F values and low p values, all models were found to be statistically significant [17]. The insignificant lack of fit achieved suggests that the models fit the experimental data correctly. The coefficient of determination was used to assess the models' accuracy (R 2 ). The significant models for S BET optimisation of produced ACs were observed as a good fit, (R 2 0.9752, R 2 0.9303, R 2 0.9673 at p<0.0001) and are represented by Eqs. (4), (5) and (6), respectively. The adjusted coefficients of determination are in reasonable agreement with the predicted coefficients of determination (Pred R 2 ). The mathematical equations in terms of actual factors obtained from the regression model predicted the responses (S BET ) and ensured the reproducibility of the experimentation. The accurate regression models fitted for S BET values of the produced ACs are quadratics.
The negative coefficient of the variables indicates its antagonistic effect on the S BET value of the produced AC. The synergetic effect of the output responses was revealed by the positive coefficients within the equations [35]. The impact of chemical activation temperature and holding time on the S BET was evaluated and displayed on the 3D surface plots of the responses (Figs. 3, 4). The evidence of a quadratic effect due to reaction temperature and holding time was demonstrated by the curvature observed on the surface plots. The impact of the two process parameters was illustrated by the slopes observed on the surface plots [36]. However, a longer residence time (135<t<180 min) reduced the S BET of the produced ACs. This might be attributed to the removal of interior carbon atoms following the highest peak of micropore formation, which breaks down the internal structure of the synthesized AC [37]. The absence of active sites such as surface oxygen groups (carbonyl and carboxyl), most likely induced by sintering of the material following the interaction of KOH and carbon, leads to surface particle adhesion. In this case, the bound particle surface flows into the pores, increasing the grain size of the material while decreasing its porosity [38,39]. As a result, previously formed micropores caused by the emission of gas products were occluded, lowering the S BET of the created ACs, which ultimately indicates that the chemical activation process parameters impact the textural properties of the generated ACs.

Optimum operating conditions
The RSM results were examined to optimize the input parameters and limitations. The purpose was to determine the best potential S BET of the produced ACs. The input parameters for the optimization of the activation process varied from 800 °C and 90 min as the lower limits to 850 °C and 180 min as the upper limits. Using the DoE's desirability function, the parameters with the highest desirability factors were chosen as the best process parameters. The results revealed that 850 °C, 147.37 min; 849.43 °C, 132.31 min; and 800 °C, 133.99 min were the optimum conditions for the activation process of HCT, HCS, and HCB, respectively. Equation 7 was used to compute the percentage of absolute error between the predicted and experimental results. Table 3 shows the error percentages obtained, which confirmed that the optimized results were consistent when compared to the experimental measurement.
According to Abdulsalam et al. [17], the best conditions of the activation process of various coals were 800 °C and a KOH:coal ratio of 4:1 to produce high S BET ACs of 1898.31, 1806.91 and 1486.74 m 2 g − 1 from the run of mines, CDs and coal slurry, respectively. Furthermore, Heidarinejad et al. [40] found that raising the temperature in the chemical activation process increases the S BET of the synthesized  AC proportionally in a review study of techniques for the synthesis and activation of AC. This is due to the accelerated thermal breakdown of materials and the elimination of volatile chemicals, which leads to the creation of porous structures. In this study, the average increase in S BET in HCT and HCS was 13% and 29%, respectively, when the temperature increased from 800 °C to 850 °C and the residence time increased from 90 min to 135 min. At 800 °C, the AC synthesized from HCB had the greatest S BET of 2243.57 m 2 g − 1 .
This optimum temperature was somewhat lower than the 850 °C needed for the optimization of AC-HCT and AC-HCS S BET . This might be attributed to the HCB's low ash content, which allows for simple interaction and quick saturation of KOH on the carbon precursor, resulting in quicker production and evaporation of gas products, leaving holes on the surface of the produced AC [38]. KOH may interact directly with the carbon atoms in the internal structure of the HCB, releasing a large amount of gas product and raising the S BET of the produced AC

Direct chemical activation
The optimal conditions of 850 °C and 135 min activation time obtained from the optimization study performed on HCT and HCS were used for the direct chemical activation of CT and CS. The BET results revealed the S BET of 1354.28 m 2 g − 1 and 1845.09 m 2 g − 1 for the produced ACs from CT and CS, respectively. However, the values of the S BET for the produced AC from direct chemical activation of coal tailings (ACT) and coal slurry (ACS) were lower than those of the synthesized AC from HCT and HCS. The S BET increases obtained using Eq. 8 were 49.29% from ACT to AC-HCT and 24.61% from ACS to AC-HCS. This suggests that the combination of HCT and the chemical activation process led to the production of ACs with higher pore volumes and larger S BET . The ACs obtained through the HTC process potentially have surface areas than a sizable proportion of ACs produced from other sources of coal used in a variety of applications such as energy storage (supercapacitor, electrode), natural gas adsorption and wastewater treatment [46][47][48][49][50].
where S BET, AC-HCx represents the specific surface of the AC from produced HC (HCx) and S BET, AC-Cx the specific surface of the AC produced from direct activation of coals (CX) ( Table 4). The obtained results corroborate the trends reported in several investigations on the carbonization pretreatment of different raw materials before the activation process with KOH reagent as indicated in Table 5.

Physicochemical and textural properties
The transformation of physicochemical properties from raw materials to optimal AC is depicted in Table 6. The HTC and Co-HTC treatments reduce the ash content of biomass  resources, but the chemical activation process significantly reduces the volatile matter content of the HCs, resulting in an increase in the carbon content of the produced ACs. It is also likely that the interaction of the activating agent and the precursor at high temperatures led to the breakdown of volatile compounds, culminating in the formation of a porous structure on the surface of the AC [69] and, as a result, lighter solid materials were produced, as also demonstrated by a decrease in carbon mass yield in Table 6. Nitrogen adsorption isotherm is the most prevalent and widely utilized method employed to assess the porous nature of ACs. The highest S BET ACs were selected for nitrogen adsorption tests at 77 K. The nitrogen adsorption isotherms for the generated ACs presented in Fig. 5a indicate a large quantity of nitrogen adsorbed on AC-HCS, followed by AC-HCB and AC-HCT, confirming that the textural properties of AC-HCS were better than the textural features of AC-HCB and AC-HCT [70]. This is consistent with the largest S BET and pore volume obtained for AC-HCS (2299.25 m 2 g − 1 and 1.96 cm 3 g − 1 ) compared to AC-HCB (2242.63 m 2 g − 1 and 1.38 cm 3 g − 1 ) and AC-HCT (2021.91 m 2 g − 1 and 1.77 cm 3 g − 1 ), respectively. The difference in textural properties between the produced ACs could be attributed to the physicochemical properties of the HCs precursors. The development of pores on the ACs is associated with the release of gas formed during the activation reaction with KOH [71]. The reaction between the KOH reagent and the HCs is equivalent to the amount of carbon content of the HCs as per the reaction in Reaction 1 [72]. Hence, the high textural parameters of AC-HCS compared to AC-HCB and AC-HCT could be assigned to the high carbon content of the HCS precursor.
(Reaction 1) 6KOH + 2C → 2K + 3H 2 + 2K 2 CO 3 . Fig. 5a show the difference in adsorption capacity of the three produced ACs. In addition, Fig. 5b shows the correlation between the specific surface area and the volume of nitrogen adsorbed (v) at a relative adsorption pressure of the nitrogen gas adsorbed. The slopes of the lines in Fig. 5b indicate that AC-HCS has the highest surface area compared to AC-HCB and AC-HCT. The nitrogen adsorption capacity of the ACs is inversely proportional to the slopes of the plotted lines [73]. The BET and BJH (Barrett-Joyner-Halenda) methods were used to determine the surface area, pore volume and average pore sizes of the selected ACs. The results revealed a parallelism between the textural properties of AC-HCS and AC-HCB, which were higher than those of AC-HCT. This is consistent with the nitrogen adsorption isotherms presented in Fig. 5. The pore size analysis revealed that the average pore diameter of the produced AC classified them in the mesopore range (2 nm<Dp<50 nm) [70,71]. The apparent density results of the products indicate that the skeleton of the high S BET AC is widely extended and occupies a larger volume compared to other ACs. This agrees with the apparent density measurement of different S BET ACs produced in many studies [17,74,75]. As reported by Jordá-Beneyto et al. [75] 6.5% (in weight) of hydrogen could be stored at 77 K and 4 MPa on AC with a minimum S BET and pore diameter of 1000 m 2 g − 1 and 20-35 Å (10 −10 m), respectively. Hence, the porous properties of the produced ACs make them suitable for hydrogen storage applications.

The adsorption isotherms in
In addition, the cumulative PSDs of the produced ACs were determined from the obtained nitrogen isotherms using the non-local density functional theory (NLDFT) method and the Barret-Joyner-Halenda method (for pore diameters above 35 Å) as shown in Fig. 6a, b. The NLDFT method normalized the pore volume to the pore width interval (differential volume dV/dW, L/Å g) to a standard slit-shaped pore model [76]. The results showed different PSD sharps for the three produced ACs. Fig. 6a reveals peaks in two zones on the PSDs plots showing a bimodal distribution in the micropore (<20 Å) and mesopore (20 Å<Dp<500 Å) regions [77]. This is consistent with the isotherm characteristics described in the previous section. The obtained peaks were identified in the micropore region for AC-HCS at widths of 16.99, 18,47 and 19.22 Å, representing pore volumes of 0.062, 0.025 and 0.023 cm 3 Å −1 g −1 , respectively. Likewise, the PSD plot of the produced AC-HCB presented peaks identified at widths of 16.79, 18,30 and 19.32 Å, corresponding to the pore volumes of 0.062, 0.047 and 0.042 cm 3 Å −1 g −1 , respectively. This confirmed that the produced AC-HCS and AC-HCB were microporous with more micropores in the AC-HCB. Therefore, the observation of the PSD reveals that the textural characteristics of the porous materials (Table 4) depend not only on the S BET [78]. Thus, a large volume of mesopores was identified in the produced  Figure 6b depicts the PSD of the produced ACs obtained by the BJH method, which follows the same pattern as the findings of the NDLFT analysis method depicted in Fig. 6a. The highest peak was found in the AC-HCS, which had the most developed porous texture. Above 35 Å in the mesopore region, peaks were also identified at widths of 48.91 Å and 78.15 Å for corresponding pore volumes of 0.0121 cm 3 Å −1 g −1 and 0.0047 for the AC-HCS. This indicates that AC-HCS had a greater mesopore proportion than AC-HCT and AC-HCB [80]. The AC-HCS had the greatest micropore structure with a greater pore volume peak, possibly leading to a higher adsorption capacity due to potential  . AC-HCS has a broader pore size dispersion than AC-HCT and AC-HCB, whereas AC-HCB has a smaller PSD. SEM images and EDS spectra of the produced ACs revealed progressively high porosity development resulting from KOH-carbon interactions (Fig. 7). The porous properties and diversity of pore sizes in the produced ACs could be observed in the SEM images obtained and agree with the PSD profiles and nitrogen adsorption isotherms. On the SEM images of AC-HCS (Fig. 7d) and AC-HCB (Fig. 7e), the heterogeneous structure with a majority of micropore sizes and randomly dispersed mesopores (bigger pores) can be clearly seen (Fig. 7g), whereas on the SEM image of AC-HCT (Fig. 7a), a comparatively uniform structure of larger and comparable mesopores was found which was also consistent with the previously described PSD analysis results. The SEM images of the produced ACs exhibit surfaces with well-developed pore patterns that lend an amorphous aspect to the material. In addition, the related EDS spectrum (Fig. 7c, f, i) shows that the produced ACs are rich in C, O, Si and Fe characteristics of the HCs precursor.

Hydrogen storage capacity of the produced ACs
There are two primary processes for hydrogen adsorption on carbon-based materials: physisorption and chemisorption. The physisorption between carbon material and hydrogen molecules is based on the van der Waals interaction. The interaction energy between the carbon material and hydrogen determines its storage capacity and may be calculated using polarization and interaction distance. Using highly polarizable materials, Mohan et al. [14] calculated the interaction energy between carbon-based material and hydrogen to be between 4 kJ mol −1 and 5 kJ mol −1 , which is smaller than the interaction energy suggested for hydrogen storage: between 10 kJ mol −1 and 50 kJ mol −1 . Consequently, carbon-based materials have a high hydrogen desorption capability. Nonetheless, Ströbel et al. [81] proved that the interaction energy may be enhanced by Kubas bindings and Spillover. Consequently, the potential hydrogen storage capacity of carbon-based materials may be determined in proportion to the specific surface area based on the greatest and lowest quantity of hydrogen adsorbed on a single graphene sheet. In chemisorption, however, every carbon atom of the material is used as an interaction site through covalent chemical bonding. This is possible at high pressure when the hydrogen molecules break apart into atoms, which favours the formation of two C-H bonds, hence shortening the distance between two neighbouring tubes and aiding the dissociative adsorption of hydrogen. As with physisorption, the high desorption capacity may be lowered at a low temperature and with the aid of a catalyst [14]. Thus, the ideal hydrogen adsorption mechanism on carbon-based materials may be defined as a two-step process beginning with the physisorption of hydrogen molecules on the surface of the material, followed by the high-pressure penetration of these molecules into the interlayer region (chemisorption). The observation of these described mechanisms indicates that the greater hydrogen adsorption is contingent on physisorption, in which the S BET plays a crucial role, before dissociative adsorption can occur. This demonstrates that the hydrogen storage capacity of carbon-based materials is closely, but not entirely dependent on their S BET . ACs might thus be used for hydrogen storage because of their high S BET and overall pore volume. These features of ACs rely, however, heavily on the feedstock and activation method.
The hydrogen storage capacity of the selected ACs with high S BET was evaluated using the adsorption-desorption isotherms. The textural properties of the ACs made them suitable for hydrogen storage applications [82]. The high-pressure volumetric method was used to measure the isotherms of hydrogen adsorption-desorption at different pressures. The results indicate that the minimum amount of hydrogen was adsorbed onto the AC-HCT, whereas the amount of hydrogen adsorbed onto the AC-HCS, and AC-HCB was 6.8% and 6.57% in weight, respectively. The validation of experimental data on hydrogen adsorption was done using the Langmuir, Toth, and Dubinin-Astakhov models (MDAs). The kinetic principle used in the Langmuir theory states that the rate of adsorption is equivalent to the rate of desorption on the porous surface and can be described by Eq. (9) in terms of fractional loading [83]. where Ɵ is the fractional loading, b is the affinity constant (gas attraction strength onto the adsorbent) and P is the pressure of adsorption. The attraction strength b can be calculated from Eq. (10): where b ∞ is the pre-exponential factor to b, Q is the isosteric heat of adsorption, R g is the gas constant and T is the adsorption temperature. The adsorption isotherms are generally expressed as the amount of adsorbed versus the adsorption pressure. Therefore, the fractional loading equation can be transformed into Eq. (11): where = C C s , and C μ and C μs are the amount of adsorbed and the saturated amount of adsorbate onto the adsorbent, respectively. μ is the adsorbed phase.
The inability to fit the adsorption data at high pressure and to describe the heterogenous nature of porous materials limit the use of the Langmuir model [14]. Hence, to address that limitation, the Toth model includes the heterogenous nature implications of the adsorbent and could be used to describe the adsorption at low and high pressure as presented in Eqs. (12)(13): where t represents the parameter that describes the heterogenous nature of the adsorbate/adsorbate system, t 0 is the parameter t at a determined temperature T 0 and α is the adsorption constant.
The simplicity and accuracy of the Toth model in the description of adsorption systems make it more suitable for the evaluation of isotherm data fitting of a couple of adsorbates into ACs [84]. However, as reported in previous studies [81,84], the fitting of physisorption data of different gases onto ACs is usually evaluated by the MDA. The MDA represented by Eq. (14) can accurately describe the equilibrium data of porous materials with high degree of heterogeneity and extended PSD.
where W represents the gas uptake in cm 3 g − 1 , W 0 shows the saturated volume uptake of gas in cm 3 g − 1 , A is the adsorption potential, E is the energy of adsorption and n is the D-A where P 0 is the saturation pressure which could be obtained from the Eq. (17): where P c and T c are the critical points of the gas at temperature T. Due to their low absolute density, a number of studies have been conducted on the potential hydrogen storage capacity of ACs derived from different precursors. These studies indicate that the hydrogen uptake value is dependent on the textural properties, chemical stability, and hydrogen adsorption process parameters of ACs [85,86]. In addition, the studies mentioned above report that hydrogen adsorption-desorption isotherms usually display little or no hysteresis and are faster at 77 K than at temperatures close to ambient temperature (273 K). Therefore, the hydrogen isotherms plots obtained in this study in Figs. 8 and 9 are in line with previous findings. The rapid kinetics of adsorption and desorption make the ACs suitable for recharging and discharging hydrogen [87]. However, the thermodynamics of the hydrogen adsorption/desorption process need to be considered for an accurate evaluation. The volumetric and gravimetric measurement results of hydrogen adsorption/desorption onto the produced ACs are represented by Figs. 8 and 9, respectively. Hydrogen adsorption on all produced ACs was found to be completely reversible. The desorption isotherms overlapped with those of adsorption without any hysteresis loop throughout the 0.1-3.5 MPa pressure range. The impact of the pressure variation emerges from the significant change in hydrogen volume adsorbed at the same temperature as shown in Fig. 9. The hydrogen adsorption/desorption results show that the hydrogen uptake (% in weight) at different temperatures increases with pressure ( Fig. 9). This is due to the increase in the interaction energy of hydrogen in the pores of the ACs [88,89]. The comparison of hydrogen isotherms for different temperatures (Figs. 8,9) shows the consistent trends of volumetric and gravimetric measurements of hydrogen adsorption/desorption onto the produced ACs. The results obtained show that the highest hydrogen uptakes (% in weight) at relatively lower pressure were observed on AC-HCB (~ 4%) compared to AC-HCS (~ 2.5%) and AC-HCT (~ 2.5% in weight), even though AC-HCS has the highest S BET . This could be attributed to the presence of many wider pores in AC-HCS, as shown in the pore size distribution (Fig. 6), which decrease the filling effect capacity of AC [90,91].
The results show that AC-HCS has the highest maximum hydrogen uptake of ~ 6.8% compared to AC-HCB Hydrogen uptake (% in weight) at (a) 77 K and (b) 273 K versus relative pressure (P/P 0 ) for various synthesized ACs. Filled and empty symbols correspond to adsorption and desorption data, respectively. AC activated carbon, AC-HCT activated carbon from coal tail-ing's hydrochar, AC-HCS activated carbon from coal slurry's hydrochar, AC-HCB activated carbon from CD (CT+CS) and SS blend's hydrochar, ads adsorption and des desorption (~ 6.57%) and AC-HCT (~ 6.12% in weight), respectively. This is consistent with the results of hydrogen adsorption on different ACs with different porous structures presented by Thomas and Ramirez-Vidal [85,90], showing that the porous structure is the major factor in determining the hydrogen adsorption capacity of ACs. Furthermore, Zhao et al. [92] showed that the functionalization of ACs can alter their hydrogen absorption capacity. The removal or donation of one electron from the surface of ACs might have a major effect on the functional group functions in hydrogen adsorption [46]. In the presence of oxygen functional groups, the interaction between adsorbate and adsorbate results in a decrease in the density of hydrogen adsorbates, which in turn reduces the maximum hydrogen absorption. The study also indicated that even at greater adsorption pressures, at ambient temperatures, only a negligible quantity of hydrogen was adsorbed. This is consistent with the isotherm data shown in Figs. 8 and 9. The low enthalpy of hydrogen adsorption is responsible for the low amount of hydrogen adsorbed at approximately ambient temperature (273 K) [91]. This indicates the temperature dependence of the hydrogen adsorption capacity of ACs. The comparison of the hydrogen uptake capacities of different ACs presented in Table 7 shows that the ACs produced in this study are among the materials with the highest hydrogen uptake capacities.   10 Process mass balance FCT filtrate from activation process of coal tailing, GCT gas from activation process of coal tailing, FCS filtrate from activation process of coal slurry, GCS gas from activation process of coal slurry, LCT liquid (process water) from the hydrothermal carbonization of coal tailing, LCs liquid (process water) from the hydrothermal carbonization of coal slurry CT coal tailing CS coal slurry, HCT coal tailing's hydrochar HCS coal slurry's hydrochar, AC-HCT activated carbon from coal tailing's hydrochar, AC-HCS activated carbon from coal slurry's hydrochar

Table 10
Comparison of the production costs of coal discards (CD) and sewage-based ACs with the production cost of various coal and sewage wastes-based ACs, as well as information on production technologies, surface area and AC production yield

Preliminary production cost and carbon emission estimations
Production costs of synthesized ACs shown in Table 8 were obtained based on the working capital estimation and mass balance shown in Fig. 10. Considering the production costs of commercial AC (Table 9), usually in the range 0.5-80 USD/kg and depending on the country considered, the synthesized AC-HCT and AC-HCS seem competitive with commercial coal-based AC, as also shown in the comparative Table 10. Additionally, the carbon retentions of synthesized ACs are shown in Table 8, indicating that the blending of SS and CDs leads to lower carbon emissions than the individual processing of CDs and SS, supporting the hydrothermal treatment's significant potential to minimize overall carbon emissions. The carbon emissions (11%-21%) reported in Table 9 are much lower than those of other thermal processes, such as conventional pyrolysis (63%) [105], microwave (70%) [105] and combustion (99%) [106].

Conclusions
H 2 storage is a major barrier to the development of a hydrogen-based energy economy, and much attention has been focused in recent decades on studies of adsorptive H 2 storage on various carbon-based adsorbents due to their low cost, good chemical, mechanical and thermal stability, easy regeneration, low densities and wide diversity of bulk and pore structures. In this paper, coal discards and SS wastes are fully characterized and transformed into value-added ACs for H 2 storage. Extensive physicochemical characterizations are carried out in order to understand the relationship between the hydrogen adsorption capacity of synthesized ACs and their physicochemical (carbon, ash, oxygen groups) and textural properties (morphology, specific surface area, pore volume, PSD). Furthermore, the physicochemical and textural properties of the produced ACs are compared to those of raw materials and achieved adsorption capacities reported in previous literature. The obtained results show that AC (AC) with the highest BET specific surface (S BET ) of 2299.25 m 2 g − 1 and 2243.57 m 2 g − 1 were obtained at 850 °C and 800 °C, respectively, using a residence time of 135 min for a KOH/HC ratio of 4:1. Using gas adsorption isotherms at 77 K, the hydrogen adsorption capacity of the produced ACs was also determined. At 35 bar, the hydrogen adsorbed onto the ACs was approximately 6.12%, 6.8% and 6.57% in weight, indicating that they could be used for hydrogen storage and sustainable carbon emissions management, as well as providing viable pathways for cost-effective energy and material circular economies for both WWTPSs and the mining industry.